COMBINING PHOTOGRAPHY AND A GEOGRAPHIC INFORMATION SYSTEM TO MEASURE WINTER BROWSE USE Roy V. Rea, Jamie D. Svendsen, and Hugues B. Massicotte Ecosystem Science and Management, University of Northern British Columbia, 3333 University Way, Prince George, British Columbia, Canada, V2N 4Z9 ABSTRACT: Browse use surveys such as the twig-length method typically used to assess browsing by ungulates are time-consuming and costly. Here, we describe a modification of the twig-length method that utilizes digital photography and a Geographic Information System (GIS) technique to quantify browse shoot removal. Linear regression analysis indicated that the cumulative shoot length (cm) and biomass removal (g) estimated with our indirect method was similar to direct measurements on Scouler’s willows (Salix scouleriana). Our results suggest that this indirect browse assessment pro- cedure could reduce field time, presumably increase sample size and efficiency, and create a photo- graphic record of each plant for long-term assessment of moose (Alces alces) browsing. ALCES VOL. 52: 67–72 (2016) Key words: browse, clipping, forage, GIS, range, survey, twig, ungulate, willow Measuring winter browse utilization of trees and shrubs by ungulates is performed by ecologists to understand ungulate diet choice and feeding requirements, and to pro- vide important information for sound range and ungulate management programs (Jensen and Urness 1981). Quantifying browse re- moval by herbivores is also fundamental to understanding the ecology of shrub and tree communities that are consumed (Bilyeu et al. 2007). Methods for determining browse use include detailed twig counts before and after ungulate browsing and percent shoot re- moval calculations (Stickney 1966, Dumont et al. 2000, Ball and Dahlgren 2002), as well as various techniques to estimate the amount of biomass removal (Bobek and Bergström 1978, Rutherford 1979, Persson et al. 2005) including broad browsed/form classifi- cation systems (Schmutz 1983, Luttmerding et al. 1990). Many of these procedures, how- ever, are time-consuming and expensive, with certain techniques requiring both fall and spring field visits in addition to mark‐ ing and tracking use of individual twigs (Jensen and Urness 1981). The twig-length method assesses utiliza- tion of shrubs and trees by measuring the amount of plant material removed by brows- ing livestock and/or wildlife (Smith and Urness 1962). Utilization is determined by measuring current annual growth on browse plants both before and after use by browsers, typically during fall and spring in temperate zones. It is an accurate and unbiased method, but has been criticized as labour intensive and requiring lengthy field time (Jensen and Scotter 1977) - criticisms commonly directed at most techniques that provide robust and accurate estimates of forage/browse use (Hyder et al. 2003, Rea et al. 2010). Digital photography has been used to simplify field counts and to provide estimates of leaf area index (Macfarlane et al. 2007), Corresponding author: Roy V. Rea, Ecosystem Science and Management, University of Northern British Columbia, 3333 University Way, Prince George, British Columbia, Canada, V2N 4Z9, reav@unbc.ca 67 mailto:reav@unbc.ca canopy closure (Guevara-Escobar et al. 2005), and fruit yields (Zaman et al. 2008). Recently, photographic methods have been tested for their accuracy and objectivity in es- timating forage use (Hyder et al. 2003) and simulated browse removal (Rea et al. 2010). In this study, we combined the use of digital photography and Geographic Information System (GIS) technology (a digital measure- ment technique) to quantify winter browse use by moose (Alces alces). Specifically, in the laboratory we combined these two approaches as an indirect method to estimate the length and biomass of willow twigs removed by simu- lated (clipping) moose browsing. We hypothe- sized that the combination of high resolution photography and digital measurements would provide accurate estimates of shoot removal and a more efficient field method to estimate browse use while maintaining the accuracy of the twig-length method. METHODS During the fall months of 2010, we removed at the stump, 50 whole saplings (~1.5 – 2.0 m tall) of Scouler’s willow (Salix scouleriana Barratt) from the forested lands surrounding the University of Northern British Columbia, Prince George, British Columbia, Canada (53.895033 °N, �122.816162 °W). The above-ground biomass of each willow sapling was weighed and photographed in the lab in front of a 10 cm lined grid (Fig. 1) using a tri-pod mounted Canon 5-D digital camera positioned at 130 cm above the ground (just above the mid-point height of our average sapling) and placed 4 m in front of the grid (Fig. 2). The camera was equipped with a wide angle lens (EF 24- 105mm Canon Zoom) set at a 50 mm focal length, high resolution, and on automatic focus and exposure. Images were framed around the midpoint of the grid so that nei- ther the camera position nor its focal length were adjusted between photographs. After photographing each plant, stems were hand-clipped at an approximate diam- eter of 4 mm (average bite diameter of local moose; Carson et al. 2007) at different inten- sities. The stem mass removed varied be- tween 3 and 86 g, and cumulative shoot A. B. C. Fig. 1. Photographs of a ~2.0 m tall willow plant before (A) and after (B) simulated browsing. Digitized shoots in (C) are marked in highlighter showing those removed by clipping. 68 BROWSE ASSESSMENTS USING GIS. – REA ET AL. ALCES VOL. 52, 2016 length between 90 and 1180 cm per plant. Plants were re-weighed and photographed a second time in front of the grid as described above. For later comparison, the total length (cm) of stems removed from each plant was measured directly with a hand ruler (near- est mm). Photographic Analysis Pre-browse and post-browse photographs were imported into ArcGIS (Version 9.3.1, ESRI 2010, Redlands, California, USA) and assessed side-by-side to identify which shoots were removed by clipping. Following cali- bration of each photograph with the cells on the measurement grid, we used ArcView’s measurement tool to overlay (see Fig. 1C) and measure the length (cm) of each shoot removed by clipping. The cumulative shoot length removed (‘browsed’) was then calcu- lated for each willow sapling. Data analysis We used linear regression analysis to de- termine 1) the relationship between the cu- mulative shoot removal measured with the simulated browsing photographic/GIS tech- nique (indirect) and the direct measurements, and 2) the relationship of biomass removal to cumulative shoot length removal associated with the indirect and direct measurement techniques. All analyses were conducted in Statistica 9.0 (Statsoft 2009). RESULTS There was a strong and significant rela- tionship (indirect shoot length (cm) = 0.991 [direct shoot length] + 2.1455; Fig. 3) be- tween shoot length estimated with the photo- graphic/GIS (indirect) technique and the direct measurements (F1,48 = 3853.9, P < 0.0001; r2 = 0.988). The relationship between biomass removal and cumulative shoot length was strong and significant for both the indirect estimates (F1,48 = 521.327, P < 0.0001; r 2 = 0.916) and direct measurements (F1,48 = 510.495, P < 0.0001; r2 = 0.914). The predict- ive relationships were similar producing nearly identical regression lines (Fig. 4): indirect shoot length = 14.247 (biomass removed) + 86.811 and direct shoot length = 14.280 (bio- mass removed) + 87.661. DISCUSSION Our indirect photographic/GIS browse assessment technique described here per- formed extremely well for estimating winter 0 200 400 600 800 1000 1200 1400 Length (cm) - Hand Measurement 0 200 400 600 800 1000 1200 1400 Le ng th ( cm ) - D ig ita l M ea su re m en t Fig. 3. The relationship (indirect shoot length (cm) = 0.991[direct shoot length] + 2.1455; r2 = 0.988) between the hand- measured (direct) and the photographic/ GIS (indirect) measurements of cumulative shoot length removal from Scouler’s wil- lows (n = 50). 4m 130cm Fig. 2. A photo-geometric representation of the protocol used to photograph willow saplings in front of our lined grid. ALCES VOL. 52, 2016 REA ET AL. – BROWSE ASSESSMENTS USING GIS. 69 shoot length and biomass removal from wil- low saplings. We expect digital photography coupled with GIS technology to be equally useful for estimating browse shoot removals from other deciduous species, although cer- tain differences may occur due to variable plant architecture (Rea et al. 2010). The method was subsequently found equally use- ful in estimating the amount and position of winter twigs removed from both Scouler’s willow and paper birch (Betula papyrifera Marsh.) by moose in cafeteria style feeding trials (Rea et al. 2015). Although tested here in a laboratory setting, the design of an easily transportable simple plastic or cloth sheet (see Schmutz 1983) with a superim- posed grid as a backdrop would provide for the photographic component to be performed in the field for pre- and post-browse utiliza- tion assessments on in situ plants (Rea et al. 2010). To test field applicability, a much longer-term study could be designed and executed where plants are photographed at the end of summer and the following spring before leaf flush. The challenge would be to mark plants such that both photographs are taken from the same perspective to en- sure an accurate estimate of browse removal. The length of shoot removal is closely correlated with biomass removal (Jensen and Scotter 1977), and our results were also strongly correlated. The indirect estimates of cumulative shoot length removal so close- ly approximated the direct measurements that we accurately estimated biomass removal using photography/GIS. We did not measure diameter where shoots were clipped (bite dia- meters), as is often done in browse surveys (Portinga and Moen 2015). The diameter at the browsed tip, however, is often predicted with regression equations developed from shoot biomass or length or vice versa, with biomass removal more typically predicted from bite diameter (Ruyle et al. 1983). In- stead, we used the length measured with GIS to predict browse biomass which, we be- lieve, circumvented the need to consider or calculate diameter at point of clipping/biting. Browse assessments in the field that em- ploy more traditional and direct measurement techniques can be costly and time consuming (Jensen and Scotter 1977). The use of this in- direct, digital technique described here could reduce the field time spent at each in situ plant by replacing direct measurements with photography. As such, there is the potential to increase the number of plants assessed in the field within a given time frame, thereby increasing sample size and presumably im- proving analytical accuracy. Digital photo- graphs taken at the time of assessment would additionally provide a permanent rec- ord of each plant that would allow for mul- tiple assessments and reduce observer bias. Databases containing such information could better describe activity, health, and popula- tion status of local moose, as well as eco- logical impacts on vegetative communities. 0 10 20 30 40 50 60 70 80 90 Biomass (g wet wt) Removed 0 200 400 600 800 1000 1200 1400 dna H dnalati gi D - ) mc( ht gneL M ea su re d HAND MEASURED GIS MEASURED Fig. 4. With respect to simulated browsing, similar relationships between shoot biomass (g wet weight) removed from willow sap- lings (n = 50) and the cumulative shoot length (cm) as measured with the direct and indirect techniques are shown. The direct relationship was: direct shoot length = 14.280 (biomass removed) + 87.661; r2 = 0.914. The indirect relationship was: indirect shoot length = 14.247 (biomass removed) + 86.811; r2 = 0.916. Note: regression lines are superimposed on one another. 70 BROWSE ASSESSMENTS USING GIS. – REA ET AL. ALCES VOL. 52, 2016 Although our technique required time spent working with the GIS to outline/digitize the twigs removed by clipping (Fig. 1C), computer algorithms and intelligent vision systems (McCarthy et al. 2010) designed to interpret the differences in plant morphology (e.g., between pre- and post-browsed plants) could reduce the time required to calculate cumulative shoot length and biomass re- moval. The creation of an artificial Cartesian coordinate grid in the GIS for registering each photograph to that grid would provide a more systematic measurement protocol. Ensuring the angle and perspective at which the pre- and post-browsed photographs are taken is critical for proper interpretation of the data. Standardized lighting, ISO, quality, and depth of field settings on the camera also need to be harmonized between pre- and post-browse photos. Like any browse assessment procedure, results will vary by species relative to plant form and twig growth characteristics. How the method performs with plants of complex architecture remains untested; however, ac- commodation for different plant forms could be approached with some resourcefulness. For example, browsed plants taller or wider than the grid could be imaged by subsections that are later summed for whole plant assess- ments. Plant and browsing height will likely define the practical limits of this technique. Nevertheless, when considering allocated field time, working within seasonal windows (e.g., assessing plants after snow melt, but before leaf flush), or attempting to increase plant numbers and data sample sizes – our technique offers an efficient and quick field method for collecting snapshots of browse use on specific plants that can be examined more closely within a controlled laboratory set- ting regardless of time and weather constraints. ACKNOWLEDEMENTS We would like to thank S. O’Keefe, T. Windsor and L. Tackaberry for their assistance with this project. We would also like to thank the staff of the Enhanced For- estry Laboratory at UNBC for providing us with the lab space and equipment necessary to complete this project. We thank S. Emmons and H. Butow for assistance with our GIS analyses. REFERENCES BALL, J. P., and J. DAHLGREN. 2002. Brows- ing damage on Pine (Pinus sylvestris and P. contorta) by a migrating moose (Alces alces) population in winter: relation to habitat composition and road barriers. Scandinavian Journal of Forest Research 17: 427–435. BILYEU, D. M., D. J. COOPER and N. T. HOBBS. 2007. Assessing impacts of large herbivores on shrubs: tests of scaling fac- tors for utilization rates from shoot-level measurements. Journal of Applied Ecol- ogy 44: 168–175. BOBEK, B., and R. BERGSTRÖM. 1978. A rapid method of browse biomass estima- tion in a forest habitat. Journal of Range Management 31: 456–458. CARSON, A. W., R. V. REA, and A. L. FREDEEN. 2007. Extent of stem dieback in trembling aspen (Populus tremuloides) as an indicator of time-since simulated browsing. Rangeland Ecology and Man- agement 60: 543–547. DUMONT, A., M. CRETE, J. P. OUELLET, J. HUOT, and J. LAMOUREUX. 2000. Popu- lation dynamics of northern white-tailed deer during mild winters: evidence of regulation by food competition. Canad- ian Journal of Zoology 78: 764–776. GUEVARA–ESCOBAR, A., J. TELLEZ, and E. Gonzalez-Sosa. 2005. Use of digital photography for analysis of canopy clos- ure. Agroforestry Systems 65: 175–185. doi: 10.1007/10457-005-0504-y. HYDER, P. W., E. L. FREDRICKSON, M. D. REMMENGA, R. E. ESTELL, R. D. PIEPER, and D. M. ANDERSON. 2003. A Digital Photographic Technique for Assessing ALCES VOL. 52, 2016 REA ET AL. – BROWSE ASSESSMENTS USING GIS. 71 Forage Utilization. Journal of Range Management 56: 140–145. JENSEN, C. H, and G. W. SCOTTER. 1977. A comparison of twig-length and browsed- twig methods of determining browse util- ization. Journal of Range Management 30: 64–67. ——, and P. J. URNESS. 1981. Establishing browse utilization from twig diameters. Journal of Range Management 34: 113–116. LUTTMERDING, H. A., D. A. DERMARCHI, E. C. Lea, D. V. MEIDINGER, and T. VOLD. 1990. Describing Ecosystems in the Field. Second Edition. MOE Manual 11. Ministry of Environment, Victoria, British Columbia, Canada. MACFARLANE, C., M. HOFFMAN, D. EAMUS, N. KERP, S. HIGGINSON, R. MCMURTRIE, and M. ADAMS. 2007. Estimation of leaf area index in eucalypt forest using digital photography. Agricultural and Forest Meteorology 143: 176–188. MCCARTHY, C. L., N. H. HANCOCK, and S. R. RAINE. 2010. Applied machine vision of plants: a review with implications for field deployment in automated farming operations. Intelligent Service Robotics 3: 209–217. PERSSON, I. L., K. DANELL, and R. BERGSTRÖM. 2005. Different moose densities and accompanied changes in tree morphology and browse production. Ecological Applications 15: 1296–1305. PORTINGA, R. L. W., and R. A. MOEN. 2015. A novel method of performing moose browse surveys. Alces 51: 107–122. REA, R.V., O. HJELJORD, and M. GILLINGHAM. 2015. Factors influencing the use of wil- low and birch by moose in winter. European Journal of Wildlife Research 61: 231–239. ——, D. P. HODDER, J. TRELENBERG, and T. M. O’BRIEN. 2010. The use of stereo- scopic photography to estimate browse use by large ungulates. Northwest Science 84: 103–108. RUTHERFORD, M. C. 1979. Plant-based tech- niques for determining available browse and browse utilization: a review. The Bo- tanical Review 45: 203–228. RUYLE, G. B., J. E. BOWNS, and A. F. SCHLUNDT. 1983. Estimating snowberry [Symphoricarpos oreophilus] utilization by sheep from twig diameter-weight rela- tions. Journal of Range Management 36: 472–474. SCHMUTZ, E. M. 1983. Browsed-class method of estimating shrub utilization. Rangeland Ecology and Management 36: 632–637. SMITH, A. D. and P. J. URNESS. 1962. Ana- lyses of the Twig Length Method of de- termining utilization of browse. Bulletin 62–9. Utah Division of Fish and Game, Salt Lake City, Utah. USA. STATSOFT. 2009. Statistica for Windows, Ver- sion 9.0. Tulsa, Oklahoma, USA. STICKNEY, P. F. 1966. Browse utilization based on percentage of twig numbers browsed. Journal of Wildlife Manage- ment 30: 204–206. ZAMAN, Q. U., A. W. SCHUMANN, D. C. PERCIVAL, and R. J. GORDON. 2008. Estimation of wild blueberry fruit yield using digital color photography. Transac- tions of the American Society of Agri‐ cultural and Biological Engineers 51: 1539–1544. 72 BROWSE ASSESSMENTS USING GIS. – REA ET AL. ALCES VOL. 52, 2016 COMBINING PHOTOGRAPHY AND A GEOGRAPHIC INFORMATION SYSTEM TO MEASURE WINTER BROWSE USE METHODS Photographic Analysis Data analysis RESULTS DISCUSSION ACKNOWLEDEMENTS REFERENCES